There is ample evidence that methane (CH4) emissions from natural wetlands exhibit large spatial variations at a field scale. However, little is known about the metabolically active methanogens mediating these differences. We explored the spatial patterns in active methanogens of summer inundated Calamagrostis angustifolia marsh with low CH4 emissions and permanently inundated Carex lasiocarpa marsh with high CH4 emissions in Sanjiang Plain, China. In C. angustifolia marsh, the addition of C-13-acetate significantly increased the CH4 production rate, and Methanosarcinaceae methanogens were found to participate in the consumption of acetate. In C. lasiocarpa marsh, there was no apparent increase in the CH4 production rate and no methanogen species were labeled with C-13. When (CO2)-C-13-H-2 was added, however, CH4 production was found to be due to Fen Cluster (Methanomicrobiales) in C. angustifolia marsh and Methanobacterium Cluster B (Methanobacteriaceae) together with Fen Cluster in C. lasiocarpa marsh. These results suggested that CH4 was produced primarily by hydrogenotrophic methanogens using substrates mainly derived from plant litter in C. lasiocarpa marsh and by both hydrogenotrophic and acetoclastic methanogens using substrates mainly derived from root exudate in C. angustifolia marsh. The significantly lower CH4 emissions measured in situ in C. angustifolia marsh was primarily due to a deficiency of substrates compared to C. lasiocarpa marsh. Therefore, we speculate that the substrate source regulates both the type of active methanogens and the CH4 production pathway and consequently contributes to the spatial variations in CH4 productions observed in these freshwater marshes.